Agentcy is a Python framework that leverages AutoGen to orchestrate a team of specialized AI agents for creative business strategy and content generation. It aims to automate complex workflows typically handled by advertising and creative agencies, providing users with plans, research, and deliverables to gain a competitive edge.

How It Works

Agentcy models a traditional agency structure, assigning distinct roles to AI agents such as Agency Manager, Researcher, Copywriter, and Strategist. These agents collaborate using the AutoGen framework, passing information and delegating tasks based on their expertise. This multi-agent approach ensures specialized handling of each project phase, from initial problem definition and research to content creation and media planning, aiming for efficient and accurate outcomes.

Quick Start & Requirements

• Install: pip install pyautogen
• Configuration: Set environment variables OAI_CONFIG_LIST or OPENAI_API_KEY, SERPAPI_API_KEY, SERPER_API_KEY, BROWSERLESS_API_KEY.
• Launch: python3 main.py
• Dependencies: OpenAI models (e.g., gpt-3.5-turbo), Serper for web search, Browserless for web scraping, Langchain for summarization.

Highlighted Details

• Employs a multi-agent system mimicking a creative agency workflow.
• Integrates tools like Serper, Browserless, and Langchain for web search, scraping, and summarization.
• Agents include specialized roles like Agency Manager, Researcher, Copywriter, and Media Planner.

Maintenance & Community

• The project is maintained by SCTY-Inc.
• The roadmap indicates plans for local LLM integration, UI development, and additional agent/tool implementations.

Licensing & Compatibility

• MIT License.
• Compatible with commercial use and closed-source linking.

Limitations & Caveats

The project currently relies on OpenAI API keys and specific third-party services, with plans to implement local LLM support in the future. Some agents are commented out, suggesting ongoing development and potential for expanded functionality.